spark load data(online&offline), the blank value in the data will become null

**Bug Description**
When using load data to import data in CSV format, the empty string in the data will become null

select into result:

**Expected Behavior**

**Relation Case**
test_select_into_load_data.yaml id:0-1. 0-2、18-2、62

**Steps to Reproduce**
```sql
create table auto_YuMSwEvO(
id int,
c1 string,
c2 smallint,
c3 int,
c4 bigint,
c5 float,
c6 double,
c7 timestamp,
c8 date,
c9 bool,
index(key=(c1),ts=c7))options(partitionnum=1,replicanum=1);
insert into auto_YuMSwEvO values
(3,'',3,22,32,1.3,2.3,1590738991000,'2020-05-03',true);
create table auto_FwREKcjN(
id int,
c1 string,
c2 smallint,
c3 int,
c4 bigint,
c5 float,
c6 double,
c7 timestamp,
c8 date,
c9 bool,
index(key=(c1),ts=c7))options(partitionnum=1,replicanum=1);
set @@SESSION.execute_mode = "online";
select * from auto_YuMSwEvO into outfile '/Users/zhaowei/code/4paradigm/OpenMLDB/auto_YuMSwEvO.csv' ;
LOAD DATA INFILE 'file:///Users/zhaowei/code/4paradigm/OpenMLDB/auto_YuMSwEvO.csv' into table auto_FwREKcjN options(mode='append');
```

The issue involves CSV data loading where empty strings become NULL. Look at the Spark load data path in OpenMLDB, likely in the Spark connector or data import modules. Check the test case referenced (test_select_into_load_data.yaml) to understand the expected behavior. Reproduce the bug using the provided SQL steps, then examine how CSV parsing handles empty values versus nulls.

Scritto dal modello di indicizzazione a partire dal testo della issue.